Advanced search
1 file | 286.00 KB Add to list

Mixed regularization method for image restoration

Ivan Cimrak (UGent) and Valdemar Melicher (UGent)
Author
Organization
Abstract
We present a new mixed regularization method for image recovery. The method is based on the combination of the bounded variation regularization and the quadratic H-k regularization. We show motivation for two distinctive terms in the energy functional, one for each regularization. We obtain rigorous results on well-posedness and stability of the underlying minimization problem. The numerical results for several case-studies give significant improvement over standard single regularizations.
Keywords
denoising, image reconstruction, Bounded variation regularization

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 286.00 KB

Citation

Please use this url to cite or link to this publication:

MLA
Cimrak, Ivan, and Valdemar Melicher. “Mixed Regularization Method for Image Restoration.” ALGORITMY 2009 : 18th Conference on Scientific Computing, Proceedings. Ed. Angela Handlovicova et al. Bratislava, Slovakia: Slovak University of Technology. Department of Telecommunications, 2009. 226–235. Print.
APA
Cimrak, Ivan, & Melicher, V. (2009). Mixed regularization method for image restoration. In Angela Handlovicova, P. Frolkovic, K. Mikula, & D. Sevcovic (Eds.), ALGORITMY 2009 : 18th Conference on Scientific Computing, Proceedings (pp. 226–235). Presented at the 18th Conference on Scientific Computing (ALGORITMY 2009), Bratislava, Slovakia: Slovak University of Technology. Department of Telecommunications.
Chicago author-date
Cimrak, Ivan, and Valdemar Melicher. 2009. “Mixed Regularization Method for Image Restoration.” In ALGORITMY 2009 : 18th Conference on Scientific Computing, Proceedings, ed. Angela Handlovicova, Peter Frolkovic, Karol Mikula, and Danial Sevcovic, 226–235. Bratislava, Slovakia: Slovak University of Technology. Department of Telecommunications.
Chicago author-date (all authors)
Cimrak, Ivan, and Valdemar Melicher. 2009. “Mixed Regularization Method for Image Restoration.” In ALGORITMY 2009 : 18th Conference on Scientific Computing, Proceedings, ed. Angela Handlovicova, Peter Frolkovic, Karol Mikula, and Danial Sevcovic, 226–235. Bratislava, Slovakia: Slovak University of Technology. Department of Telecommunications.
Vancouver
1.
Cimrak I, Melicher V. Mixed regularization method for image restoration. In: Handlovicova A, Frolkovic P, Mikula K, Sevcovic D, editors. ALGORITMY 2009 : 18th Conference on Scientific Computing, Proceedings. Bratislava, Slovakia: Slovak University of Technology. Department of Telecommunications; 2009. p. 226–35.
IEEE
[1]
I. Cimrak and V. Melicher, “Mixed regularization method for image restoration,” in ALGORITMY 2009 : 18th Conference on Scientific Computing, Proceedings, Podbanske, Slovakia, 2009, pp. 226–235.
@inproceedings{533123,
  abstract     = {{We present a new mixed regularization method for image recovery. The method is based on the combination of the bounded variation regularization and the quadratic H-k regularization. We show motivation for two distinctive terms in the energy functional, one for each regularization. We obtain rigorous results on well-posedness and stability of the underlying minimization problem. The numerical results for several case-studies give significant improvement over standard single regularizations.}},
  author       = {{Cimrak, Ivan and Melicher, Valdemar}},
  booktitle    = {{ALGORITMY 2009 : 18th Conference on Scientific Computing, Proceedings}},
  editor       = {{Handlovicova, Angela and Frolkovic, Peter and Mikula, Karol and Sevcovic, Danial}},
  isbn         = {{9788022730327}},
  keywords     = {{denoising,image reconstruction,Bounded variation regularization}},
  language     = {{eng}},
  location     = {{Podbanske, Slovakia}},
  pages        = {{226--235}},
  publisher    = {{Slovak University of Technology. Department of Telecommunications}},
  title        = {{Mixed regularization method for image restoration}},
  year         = {{2009}},
}

Web of Science
Times cited: